Answers are grounded in this paper and cite their sources.
Brushless direct current (BLDC) motors are employed for electric vehicles (EVs), medical equipment, and military robots because of their fast dynamic speed response and wide range of speed regulation characteristics. However, conventional methods are not suitable for dynamic speed-varying conditions, which leads to parameter variation issues and longer computational time. Therefore, in this article, a novel Extended Kalman Filter (EKF) for sensorless speed detection of a BLDC motor is implemented using a machine learning-based physics-informed neural network (PINN) algorithm. Variables obtained from the PINN training are incorporated into the novel EKF; then, with the implemented model, the BLDC motor speed tracks the reference speed in dynamic, speed-varying conditions, thereby eliminating parameter variation issues and achieving shorter computational time as well. The implemented sensorless speed estimation scheme is compared with the conventional algorithms, and the results are presented using LAUNCHXL-F28379D. Also, compared with other conventional sensorless speed estimation techniques, the implemented method achieved reduced speed errors of 0.003% and 0.03% at 1500 rpm and 2000 rpm, respectively.
No takes yet. Share an insight, caveat, or question.
Sasidharan et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: